Hi all,
We have been having an offline discussion of this issue. I am posting
the correspondence so that it can be referenced in the future. Thanks.
-karl
Sebastien,
Nope, you are doing it exactly right. I fixed that problem a while ago
but apparently neglected to upload my code changes to Bioconductor...
In the mean time here is the fixed function for you to use:
aggregateExprs <-
function(x,package="hgu133plus2",using="ENTREZID",FUN,...) {
if(class(x) != "matrix" && !is(x, "ExpressionSet"))
stop("need matrix or ExpressionSet")
if(is.null(package) && is(x, "ExpressionSet"))
package <- annotation(x)
if(is.null(package))
stop("annotation package name is required")
if(!require(package,character.only=TRUE))
stop(package," is not available")
pPos <- paste("package",package,sep=":")
if(grep(".*db",package)) package <- gsub(".db","",package)
nEnv <- paste(package,using,sep="")
Env <- get(nEnv,pos=pPos)
if(is(x, "ExpressionSet")) {
ids <- featureNames(x)
x <- exprs(x)
} else {
ids <- rownames(x)
}
lls <- mget(ids,env=Env,ifnotfound=NA)
if(length(lls)!=length(unlist(lls))) for(i in 1:length(lls))
lls[[i]] <- lls[[i]][1]
lls <- unlist(lls)
f <- factor(lls)
undupx <- aggregate(x,by=list(f),FUN,...)
rownames(undupx) <- as.character(undupx[,1])
undupx <- as.matrix(undupx[,-1])
return(undupx)
}
-karl
-----Original Message-----
From: Sebastien Gerega [mailto:sgerega@gmail.com] On Behalf Of
Sebastien Gerega
Sent: Wednesday, July 09, 2008 7:17 PM
To: Dykema, Karl
Subject: Re: [BioC] ENTREZ identifiers for PGSEA
Hi,
thanks for your reply. I have just had a quick go using the following
command:
aggregateExprs(exprs(eset), "lumiHumanAll", FUN=mean, na.rm=TRUE)
but I get the error:
"There is no package called "lumiHumanAll"
If I try specifying the library as "lumiHumanAll.db" then I get the
error:
variable "lumiHumanAll.dbENTREZID" was not found...
Am I doing something wrong?
thanks,
Sebastien
Dykema, Karl wrote:
> Sebastien,
>
> You should take a look at our function "aggregateExprs". As long as
you have an annotation environment for your lumi microarray, it is
quite simple. Let me know if you run into any problems.
>
> -karl
>
>
>
> ------ Forwarded Message
> From: Sebastien Gerega <seb at="" gerega.net="">
> Date: Wed, 09 Jul 2008 16:31:29 +1000
> To: "bioconductor at stat.math.ethz.ch" <bioconductor at="" stat.math.ethz.ch="">
> Subject: [BioC] ENTREZ identifiers for PGSEA
>
> Hi,
> I would like to use a lumi expression set with the PGSEA package and
GO
> to generate genesets. I understand that the identifiers used in the
data
> matrix need to be the same as those used in the genesets. Therefore,
> since I am using the GO database I need to use the ENTREZ
identifiers
> for my matrix. How can I go about doing this? Obviously I cannot
have
> multiple rows with the same name so I will have to remove duplicates
and
> should probably do so based on variance or something. Is there an
> elegant way to do what I need?
> thanks,
> Sebastien
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
>
https://stat.ethz.ch/mailman/listinfo/bioconductor
This email message, including any attachments, is for
th...{{dropped:6}}
